Place Matters: Why the Location of Low-Income Housing Projects is Important to Neighborhood Revitalization
Housing projects assisted by the U.S. Department of Housing and Urban Development (HUD) can significantly impact the neighborhoods in which they are located. Some have been well maintained and are regarded as assets, offering badly needed affordable housing in their communities. Others, with deteriorated structures and large concentrations of poor tenants can create negative spillover effects for their surroundings. Recently, HUD has begun to see these projects as a policy opportunity. In its Choice Neighborhoods initiative, HUD is, for the first time, using the redevelopment of some of these projects as a springboard for the broader revitalization of the neighborhoods around them.
Whatever the condition or potential of these projects, their location is critical to informing public policy. This commentary offers an overall summary of the distribution of HUD-assisted, project-based housing units within and among the 100 largest metropolitan areas as of 2008 (this category includes public housing projects and privately owned projects that provide subsidies under assistance contracts with HUD). Knowing how these units are distributed provides some insight into which metros and neighborhoods will be most affected by HUD’s broader revitalization goals.
In 2008, the number of occupied project-based housing units totaled 2.1 million nationally, with 76 percent of them (1.6 million) in the 100 largest metros. HUD project assistance is unevenly distributed geographically, as shown in these charts.
The first chart shows that project-based assistance is concentrated in low-income neighborhoods (census tracts with poverty rates of 20 percent or more). Only one-quarter of all census tracts in the 100 metros qualify as low income, but they account for 6 of every 10 project-based units (total of 950,000). Only 5 percent of all tracts are in the extreme poverty category (those with a poverty rate of 40 percent or more), but are home to 20 percent of all project-based units. Still, the extent of HUD-assisted projects in moderate- and higher-income neighborhoods (those with poverty less than 20 percent) is far from trivial-654,000 units accounting for 41 percent of all project-based units.
The next two charts show the variation in clustering (where project-based assistance represents a high share of all housing in the neighborhood). Among low-income tracts, we find, at one extreme, 800 tracts where clustering is substantial. In these tracts, HUD projects (with 441,700 units) are a dominating presence, accounting for 30 percent or more of all units. What happens to these projects, for good or ill, will have a powerful influence on the future of their neighborhoods overall.
In 1,400 tracts, project-based units represent from 10 to 30 percent of the total housing (not so dominant but still influential). HUD-assisted projects account for less than 10 percent of all units in 2,000 tracts. Most noteworthy, however, is that 5,500 tracts (well over half of all low-income tracts) have no HUD project-based units at all, meaning that HUD projects will not be a factor in neighborhood revitalization.
Among moderate- and high-income neighborhoods (those with poverty rates below 20 percent) HUD-assisted projects are less prevalent: 81 percent of the tracts do not have any project-based units and in most others (15 percent), such units represent less than 10 percent of the housing stock.
We also found that the share of HUD-assisted projects varies greatly among metropolitan areas. At the high end, these projects accounted for 4.5 percent or more of all housing in metropolitan Providence, New York, St. Louis, Seattle, and Boston. The absolute size of metropolitan New York’s project-based stock stands out-with 357,000 occupied units, New York has more than four times the total of its nearest rival (Chicago). At the low end, project-based units accounted for less than 0.8 percent of all units in metropolitan Dallas; Riverside-San Bernardino, California; Boise City, Idaho; Phoenix; and Bradenton, Florida.
This distribution of project-based assistance is poorly correlated with metropolitan poverty rates. Of the 100 metros we studied, five of the top 20 in terms of poverty ranked among the lowest 20 in terms of project-based assistance: Bakersfield, California; San Antonio; Tucson; Albuquerque; and Houston. This result is not explained by HUD allocations policies. Differences in the degree to which developers and housing authorities sought HUD projects were most important; some cities (more often in the snowbelt) were aggressive in seeking this form of assistance, while others (often in the sunbelt) generally avoided it. Performance also played a role; HUD reallocated funds from cities unable to complete approved projects to those with better production capacity.
Metros also vary widely in the densities of project-based units in low-income neighborhoods. Density is a useful measure because it indicates the degree to which HUD projects are likely to be important to neighborhood revitalization planning (which focuses on low-income neighborhoods). Metropolitan New York tops this list with 20 HUD project units per 100 total units in low-income tracts. Metropolitan Boston; Honolulu; Washington, DC; and Worcester, Massachusetts come next with 15 to 19 project-based units per 100. At the other end of the spectrum are five metros with two or fewer HUD project units per 100: Riverside-San Bernardino; Bakersfield, California; McAllen, Texas; Boise City; and Scranton, Pennsylvania.
Send us your comments to help further the discussion.